Transcript Research on Macrofamilies: The States of the Art

Research on Macrofamilies:
The States of the Art
Bernard Comrie
Max Planck Institute for Evolutionary
Anthropology and University of California
Santa Barbara
[email protected]
African Macrofamilies
•
•
•
•
Afroasiatic
Niger-Congo
Nilo-Saharan
Khoisan
Niger-Congo
• Mande - inclusion questionable
• Kordofanian - inclusion of Kadugli
widely rejected, may be Nilo-Saharan
• Atlantic - may not be a unit, some parts
may not be Niger-Congo
• Bulk of phylum generally accepted
Nilo-Saharan
• Lower-level groupings generally
accepted, higher-level groupings
questionable
• Songhai - almost universally excluded
Khoisan
• Hadza - no clear relation
• Sandawe - possibly related to Central
• South African Khoisan
– Central - no clear relation to N, S
– Northern - no clear relation to Southern
– Southern - no clear relation to Northern
– ≠Hõa may be related to Northern
Family Tree Problems
• Presupposes “parthenogenesis”
• Areal spread of innovations (“Wave
theory”)
• Loans from other languages
Wave Theory
•
•
•
•
•
•
“Rhenish Fan” in Germany
North:
maken dorp dat
Cologne:
machen dorp dat
Koblenz:
machen dorf dat
Frankfurt:
machen dorf das
South:
machen dorf das
appel
appel
appel
appel
apfel
• Indo-Iranian: dative plural in bh; shift of
*kj to s
• Balto-Slavic: dative plural in m; shift of
*kj to s
• Germanic: dative plural in m; no shift of
*kj to s
Multiple Origins / Language
Contact
• Possible definition of genealogical
relatedness of languages: Two
languages descend from a common
ancestor if a substantial portion of basic
vocabulary, (inflectional) morphology (if
present), and syntax descend from that
ancestor.
• Frequent, but usually tacit assumption,
that there is a hierarchy:
Morphology > Basic vocabulary > Syntax
Cf. Comrie on Haruai (in relation to
Hagahai and Kobon), comparing
morphology and basic vocabulary
Basic Vocabulary Loans
• Thai càmùuk ‘nose’ < Khmer crɑmoh,
cf. forms like daŋ, laŋ, naŋ in other Tai
languages
• English they, them, their <Scandinavian
Morphological Loans
• English loans from Latin / Greek
retaining original number morphology:
– criterion
– crisis
– syllabus
– formula
criteria
crises
syllabi / syllabuses
formulae / formulas
• Romani singular / plural morphology
– Inherited: kher, PL kher-a ‘house’; šer-o, PL šer-e
‘head’; no plural in -i
– Early Greek loans introduce type for-os (in many
varieties > for-o), PL for-i ‘town’
– Other contact languages introduce other plural
markers, e.g. -uri < Rumanian, even with older
words (some varieties have for-uri)
• Copper Island (Mednyj) Aleut has Aleut
basic vocabulary and nominal
morphology, but Russian verb
morphology, even for inherited verbs:
aba-ju ‘I work’
aba-im ‘we work’
aba-iš ‘you work’
aba-iti ‘you (PL) work’
aba-it ‘s/he works’ aba-jut ‘they work’
Syntactic Loans
• Takia has Austronesian vocabulary and
morphology, but the same grammatical
structure as its Papuan neighbor
Waskia.
• Haitian Creole has French vocabulary
and either West African or “universal”
syntax.
Regular Sound
Correspondences
• Important:
Traditional historical linguists
Starostin / Russian school
• Unimportant:
Greenberg / Ruhlen
Opaque but Regular
Correspondences
• French fils /fis/, Spanish hijo /'ixo/ ‘son’
• German fünf, Russian pjatj, Armenian
hing ‘five’ < Proto-Indo-European
*penkwe
• Mbabaram dog ‘dog’ < Proto-Australian
*gudaga
Exceptions to the Regularity of
Sound Change
• Latin quinque ‘five’ for expected *pinque
• French cinq /sɛk̃ / for expected /kɛk̃ /
• English she /šiː/ < Old English sēo for
expected */siː/; probable developments:
siːo > sjoː > šoː > šuː (dialect form)
> siːe > sjeː > šeː > šiː
Explanations for Similarities
•
•
•
•
Common ancestry
Borrowing
Naturalness (e.g. onomatopoeia)
Chance
Eliminating Chance
• English
German
•
•
•
•
•
Arm
Finger
Hand
Lippe
Nase
arm
finger
hand
lip
nose
y
y
y
y
y
• English
German
•
•
•
•
•
Finger
Hand
Lippe
Nase
Arm
arm
finger
hand
lip
nose
n
n
n
n
n
Loanword Typology Project
• To study patterns of lexical borrowing
across a number of languages from
different language families, to ascertain
in particular if certain lexical items
(relating to particular semantic fields, or
particularly basic items) are less prone
to borrowing than others.
Some Successes
• “Borderline” convincing cases
– Altaic (or at least some branches thereof)
– Indo-European and Uralic
– (Narrow) Trans-New Guinea
• Yeniseic (western Siberia) and Na-Dene
(northwestern North America)
– a few dozen possible plausible cognates
– very similar, complex, unusual verb
morphology, though similarities could be
typological (?contact rather than common
ancestor)
(work by Edward Vajda)